摘要
应用人工神经网络滚动预测方法，采用5-25-5神经网络结构建立实验区红桦林5个龄组的蓄积量预测模型，预测了实验区2000-2004年红桦各龄组蓄积量。幼龄林预测平均相对误差为1．138％，中龄林预测平均相对误差为0．051％，近熟林预测平均相对误差为0．362％，成熟林预测平均相对误差为0．156％，过熟林预测平均相对误差为0．009％。实验的预测值与实际值吻合度高，用该模型进行预测的可信度是较高的。
A prediction model was establishing based on 5-25-5 Artificial Neural Network （ANN＇） structure, five-age group wood volume of Betula albo-sinensis of experimental forests from 2000 to 2004. The results showed that the average relative errors of younger aged group, middle aged group, nearly mature group, mature group and over mature group were 1. 138 04%,0. 050 8%,0. 362 4%,0. 155 8% and 0. 008 7 % respectively. This suggested that the predicting numbers fit the factual numbers well and the model was simple, practicable and precise.
出处
《西北林学院学报》
CSCD
北大核心
2006年第1期178-182,共5页
Journal of Northwest Forestry University
